Doubly robust estimation of attributable fractions in survival analysis

Sjölander, Arvid; Vansteelandt, Stijn

doi:10.1177/0962280214564003

Cited by 36 publications

(30 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To sum up, we proposed two novel estimands of the PAF to quantify the burden of time‐dependent harmful exposures on a population level. The two estimands also account for competing risks, filling a gap in the literature (Sjölander & Vansteelandt, ). Finally, our innovative way of using dynamic prediction and landmarking allows for adjustment of time‐varying confounding in a straightforward way.…”

Section: Discussionmentioning

confidence: 99%

The population‐attributable fraction for time‐dependent exposures using dynamic prediction and landmarking

et al. 2019

View full text Add to dashboard Cite

The public health impact of a harmful exposure can be quantified by the populationattributable fraction (PAF). The PAF describes the attributable risk due to an exposure and is often interpreted as the proportion of preventable cases if the exposure was extinct. Difficulties in the definition and interpretation of the PAF arise when the exposure of interest depends on time. Then, the definition of exposed and unexposed individuals is not straightforward. We propose dynamic prediction and landmarking to define and estimate a PAF in this data situation. Two estimands are discussed which are based on two hypothetical interventions that could prevent the exposure in different ways. Considering the first estimand, at each landmark the estimation problem is reduced to a time-independent setting. Then, estimation is simply performed by using a generalized-linear model accounting for the current exposure state and further (time-varying) covariates. The second estimand is based on counterfactual outcomes, estimation can be performed using pseudo-values or inverse-probability weights. The approach is explored in a simulation study and applied on two data examples. First, we study a large French database of intensive care unit patients to estimate the population-benefit of a pathogen-specific intervention that could prevent ventilator-associated pneumonia caused by the pathogen Pseudomonas aeruginosa. Moreover, we quantify the population-attributable burden of locoregional and distant recurrence in breast cancer patients. K E Y W O R D Sattributable risk, competing risks, dynamic prediction, landmarking, time-dependent exposure Biometrical Journal. 2020;62:583-597.

show abstract

Section: Discussionmentioning

confidence: 99%

The population‐attributable fraction for time‐dependent exposures using dynamic prediction and landmarking

et al. 2019

View full text Add to dashboard Cite

show abstract

“…PAF was quantified for all three relationships and was defined as the proportion of cases that could be prevented if the risk factor would be completely removed from the population. PAF was based on Cox models adjusted for confounding factors using the R package ‘AF’ (V.0.1.4) 19. In order to calculate PAF, non-HDL-c was dichotomised at below or above 2.6 mmol/L, for SBP a cut-off at 140 mm Hg was used, and smoking was analysed as current smoking versus never or former smoking.…”

Section: Methodsmentioning

confidence: 99%

Cardiovascular risk factors and the risk of major adverse limb events in patients with symptomatic cardiovascular disease

Hageman

Borst

Dorresteijn

et al. 2020

Heart

View full text Add to dashboard Cite

AimTo determine the relationship between non-high-density lipoprotein cholesterol (non-HDL-c), systolic blood pressure (SBP) and smoking and the risk of major adverse limb events (MALE) and the combination with major adverse cardiovascular events (MALE/MACE) in patients with symptomatic vascular disease.MethodsPatients with symptomatic vascular disease from the Utrecht Cardiovascular Cohort - Secondary Manifestations of ARTerial disease (1996–2017) study were included. The effects of non-HDL-c, SBP and smoking on the risk of MALE were analysed with Cox proportional hazard models stratified for presence of peripheral artery disease (PAD). MALE was defined as major amputation, peripheral revascularisation or thrombolysis in the lower limb.ResultsIn 8139 patients (median follow-up 7.8 years, IQR 4.0–11.8), 577 MALE (8.7 per 1000 person-years) and 1933 MALE/MACE were observed (29.1 per 1000 person-years). In patients with PAD there was no relation between non-HDL-c and MALE, and in patients with coronary artery disease (CAD), cerebrovascular disease (CVD) or abdominal aortic aneurysm (AAA) the risk of MALE was higher per 1 mmol/L non-HDL-c (HR 1.14, 95% CI 1.01 to 1.29). Per 10 mm Hg SBP, the risk of MALE was higher in patients with PAD (HR 1.06, 95% CI 1.01 to 1.12) and in patients with CVD/CAD/AAA (HR 1.15, 95% CI 1.08 to 1.22). The risk of MALE was higher in smokers with PAD (HR 1.45, 95% CI 0.97 to 2.14) and CAD/CVD/AAA (HR 7.08, 95% CI 3.99 to 12.57).ConclusionsThe risk of MALE and MALE/MACE in patients with symptomatic vascular disease differs according to vascular disease location and is associated with non-HDL-c, SBP and smoking. These findings confirm the importance of MALE as an outcome and underline the importance of risk factor management in patients with vascular disease.

show abstract

“…The AF was used to quantify the public health impact of FI on the survival outcome in the cohort of single responders based on a time-constant survival model [10][11][12]. The AF is an integrated predictive measure that takes into account both the prevalence of a risk factor in the population as well as the strength of association between risk factor and outcome.…”

Section: Methodsmentioning

confidence: 99%

“…Secondly, we estimate FI-mortality associations in a time-dependent manner to elucidate the predictive value of FI among the young. Lastly, we investigate the number of deaths that could be attributable to increased population levels of frailty by estimating attributable fractions (AFs) [10][11][12].…”

Section: Introductionmentioning

confidence: 99%

The frailty index is a predictor of cause-specific mortality independent of familial effects from midlife onwards

Ploner

Karlsson

et al. 2018

Preprint

View full text Add to dashboard Cite

Background. Frailty index (FI) is a well-established predictor of all-cause mortality, but less is known for cause-specific mortality and whether familial effects influence the associations.Furthermore, the population mortality impact of frailty remains understudied.Objectives. To estimate the predictive value of frailty for all-cause and cause-specific mortality, and to test whether the associations are time-dependent. We also assessed the proportion of deaths that are attributable to increased levels of frailty.Methods. We analyzed 42,953 participants from the Screening Across the Lifespan Twin Study (aged 41-95 years at baseline) with up to 20-years' mortality follow-up. The FI was constructed using 44 health-related items. Deaths due to cardiovascular disease (CVD), respiratory-related causes and cancer were considered in the cause-specific analysis. Generalized survival models were used in the analysis. Results.Increased FI was associated with higher risks of all-cause, CVD, and respiratory-related mortality. No significant associations were observed for cancer mortality. No attenuation of the mortality associations was found in unrelated individuals when adjusting for familial effects in twin pairs. The associations were time-dependent with relatively greater effects observed in younger ages. The proportion of deaths attributable to FI levels >0.10 were 13.0% of all-cause deaths, 14.7% of CVD deaths and 12.5% of respiratory-related deaths in men, and 12.2% of allcause deaths, 9.9% of CVD deaths and 21.9% of respiratory-related deaths in women. Conclusions.Increased FI predicts higher risks of all-cause, CVD, and respiratory-related mortality independent of familial effects. Increased FI levels have a significant population mortality impact in both men and women.

show abstract

Doubly robust estimation of attributable fractions in survival analysis

Cited by 36 publications

References 18 publications

The population‐attributable fraction for time‐dependent exposures using dynamic prediction and landmarking

The population‐attributable fraction for time‐dependent exposures using dynamic prediction and landmarking

Cardiovascular risk factors and the risk of major adverse limb events in patients with symptomatic cardiovascular disease

The frailty index is a predictor of cause-specific mortality independent of familial effects from midlife onwards

Contact Info

Product

Resources

About